The Photolysis of Azomethane. III. The Effect of Nitric Oxide and the

California, Berkeley, Calif. (4) Cf. (a) F. ... represented by reaction 1 as well as by reaction 2. Unfortunately ... ticularly at low temperatures (~...
4 downloads 0 Views 944KB Size
2. During the rcarrangeineiit of the catechol esters small amounts of 3-acylcatecholsare formed. 3. The -&acylcatechols ma)’ be converted into the corresponding alkvlcatechols. .

COTrRIFlT TIOY FROM THE

The Photolysis of Azornethane. BY

4. The alkylcatechols are effective germicides. The phenol coefficient increases with the length of the alkyl chain. GLENOLDEN, P n NNA . - - --_-__

DEPARTMENT O F CHEMISTRY

RECEIVED AUGUST10, 1937

OF

NEWYORK

UNIVERSITY]

III. The EEect of Nitric Oxide and the Nature of the Primary Step1t2

THOMAS w. DAVIS,FRANCIS P. JAHN

AND MILTON BURTON3

+

CH, A ----f CHsA (6) Hitherto it has been tacitly assumed*that azomethane decomposes almost entirely to form free where A is used to symbolize a molecule of azomethyl radicals and molecular nitrogen in one methane. The product is itself a free radical.8 primary act. Patat and Sachsse,” among others, The results obtained were consistent with the have considered the possibility of a primary step hypothesis that part of the primary step might be yielding N=NCH3 and CH3. The S=NCH3 represented by reaction 1 as well as by reaction 2. radical has been assumed to be very unstable;4a3636Unfortunately, the evidence on this point was not although Patat and SachsseSahave indicated the conclusive. implications of an assumed stability, Patatsb has One of usJdiscovered in a study of the pyrolysis been inclined in favor of the assumption of the of azomethane that nitric oxide almost cominitial split into free radicals. In general, how- pletely inhibits any pressure change in the early ever, there has been no evidence for a valid con- stages of the reaction. Although such an obserclusion as to the relative probabilities of the paths vation is not directly interpretable1* the result suggested the possibility of making use of a similar CH?N=NCHT + IZP CJHG+ (1) method for a direct determination of the nature of CHBN=NCHa f h V +3CH1 + Nl (2) C€I,&-NCH t lit, -+ CH? + the primary reaction in the photolysis.l L Staveley and Hinshehood’2 have made use of Itiasmuch as azomethane has been used as a sourcc the assumption that nitric oxide reacts very of free methyl radicals in estimating chaiii rapidly with free radicals ixi order to detect the Iength~,~“ it is evidently important to have inforparticipation of radicals in reaction chains. 111 mation as to the primary step. the work described below we have made use of the Burton, Davis and Taylor6,’ have shown that same assumption in determining the nature of the methyl radicals formed in the photolysis of azoprimary step in the photolysis of azomethane. methane do not disappear exclusively by the -4lthough we have attempted to avoid the pitfalls reaction involved in kinetic studies dependent on pressure ZCH, SI -+ CzHa + XI (1 r,) changes by studying the products of the photolybut that some probably disappear by an associa- ses analytically, it must be emphasized that our tion reaction of the type conclusions are based upoii the assumption that Sj

+

(1) Abstract from .i t 1 i l - w to be m b m t t e d by Francir P Iahn t o the Graduate School of New York University in partial fulfilment of the requirements for the degree of Doctor of Philosophy (2) Presented before the Division of Physical and Inorganic Chemistry a t the Rochester Meeting of the American Chemical Society, Sept. 6, 1937. (3) Present address Department of Chemistrq l m i x ersitv of California, Berkeley, Calif (4) Cf. (a) F. 0. Rice and R. K. Rice, “The Aliphatic Free Radicals.” The Johns Hopkins Press, Baltimore, Md , 1Q.35, p. 140, (b) 0. K Rlce and Sickman, J . Chem. Phys , 4, 242 (1‘136), (4Allen and Sickman, TWIS JaURI*&, Bd, 2031 (1934). (5) (a) Patat and Sachsx, 2. physsk. Chem., B31, 108 (1935J, (b) Patat, Nalwwissenschaften, 43, 801 (1935): Nachr. Ges Wiss G%#finge:cn,Mafh.-Dhysik Klossc, Fachgruppeci 11, IN. F.] %,77 (1936) (6) Burton, Dapis and Taylor, THISJOUXXAL, 69, 1038 (1937) ( 7 ) Burton, T).r\isai~dTaylor, ihid ,69, I‘lhO (10371

(8) The comparatively low probability of reaction 15 ib well illustrated in the work on the alkyl halides (cf. Bonhoeffer a n d Harteck, “Grundlagen der Photochemie,” Verlag von Theodor Steiukopff, Dresden, 1933, p. 172) and in the work on acetone [Spence and WiId, J . Chem. Soc , 352 (1987)l. In the former case the yield of ethane is negligible; in the latter, analyses show that there is a high probability that methyl disappears by other reactions, par ticularly a t l o w temperatures (--Zoo) (‘1) Jahn, unpublished work. (10) I‘he work cited in references 6 and 7 indicates that failure to observe a pressure increase may be due t o a multifold effect involving reactions such as 1, 2, 6, 16 and 18. Cf Riblett and Rubin. THISJOURNAL, 69, 1537 (1937). (11) Cf (a) Linnett and Thompson, Trons. Faraday Soc , 33, 501 (1937), (b) Thompson and Linnett. abrd , 33, 874 (19371 112) Staveley and Hinshelwood, Proc Roy Soc (London) 8 1 6 4 , 1.3; (10 Hi), V a / i i w , 13TyI, 2‘1 (1‘136) 1. ChPm %If , 812 (lQ36)

Jan., 1938

NITRICOXIDEIN

THE

nitric oxide reacts only with free radicals in the system studied. Our results indicate that azomethane decomposes almost exclusively by reaction 2, that reaction 6 takes place quite readily and that ethane formed in the photolysis of azomethane probably does not result from reaction 15.

Experimental Method The procedure employed was similar to that described by Burton, Davis and T a y l ~ r . ~ J Azomethane.-The method of preparation of azomethane was slightly modified. On theoretical grounds it appeared advisable to add the usual potassium chromate solution to the iced solution of dimethylhydrazine hydrochloride. It seemed to us that in this way the danger of formation of higher products of oxidation ( e . g., methyl nitrate) and the danger of explosion13 both would be reduced and that the yield of azomethane would be correspondingly increased. We have found that the sole advantage of this procedure is that foaming is prevented until all the dimethylhydrazine hydrochloride has reacted. The yield is, as a matter of fact, reduced and a difficulty arises through the formation of a large quantity of lowboiling impurity (perhaps methyl nitrite) which is not removed conveniently. Our procedure has been to use the method of purification already described6 and finally to use only the last fractions of evaporation of the condensed liquid. Samples so prepared had a v. p. of -752 mm. a t 0’ and were considered satisfactory for use. Nitric Oxide.-Nitric oxide was prepared by dropping mercury into a 25% solution of nitrosylsulfuric acid in sulfuric acid, prepared according to the method described by Biltz and Biltz.14 Precautions were taken to remove residual air by repeated evacuation of the acid and by flushing out the system with nitric oxide before a sample was considered satisfactory for use. Analytical Method.-The analytical method was modified to ensure the removal of residual nitric oxide preliminary to a determination of the reaction products. This usually was done by the introduction of oxygen before treatment of the product gases (as already described6) with bromine and potassium hydroxide. No effect was ever observed on subsequent absorption in potassium hydroxide; the oxygen added was removed quantitatively by chromous chloride and pyrogallol treatments.8 Evidently any residual nitric oxide was always removed by the analytical treatment before this point of the procedure. Procedure and Results.-The general procedure of the runs was similar to that already described.eJ The reaction vessel was opened to the nitric oxide reservoir and filled to the desired pressure. It was then shut off from the reservoir and the intermediate tubing was exhausted. Azomethane was then introduced slowly through a very slightly opened stopcock until the desired pressure was attained. In preliminary experiments it was established (13) Cf.Jahn, THIS JOURNAL, 69,1761 (1937). (14) H. Biltz and W. Biltz “Laboratory Methods of Inorganic Chemistry,” translated by Hall and Blancbard, John Wiley and Sans. Inc.. New York, 1909,p. 204.

PHOTOLYSISOF AZOMETHANE

11

that in a few minutes the mixing was sufficient to yield the same results as after several hours of standing. The usual time was about five minutes. In the work around 20’ we used a water thermostat with iilter and a t higher temperatures an air thermostat with filter, the same as those already described. The Corning No. 534 blue nultra flter, transmitting only wave lengths longer than 3500 A., was used in all the experiments except one. In experiment 19, as indicated in Table I, no filter was used. The thickness of Pyrex glass in that experiment transmits down to >2900 A. The runs were started by exposing the reaction vessel to the light from the constricted mercury arc16 and stopped by interrupting the arc. The current in the arc was held at -4.2 amperes during runs. Since the arc was not kepi perfectly steady and since the arc distance in the 20’ runs was much less than in the higher temperature experiments, no direct comparison of the precise slopes of the rate curves can be made. The general shapes of the curves, however, are interesting and will be discussed later. The samples of gas were removed for analysis immediately after the conclusion of the runs and treated as previously described. The results of the analyses are summarized in Table I and the graphs of four typical runs are shown in Fig. 1, in which A P is the pressure change observed a t any time.

5 1 - 7

I

I

1

!

-

I

13

- 101

I

20

I

I

I

I

I

I

60 80 100 Time, minutes. Fig. 1.-The course of pressure change in several experiments; the dotted portion of the curve for expt. 10 represents the pressure change after illumination had ceased. 0

40

In the table, PKO is the initial pressure of nitric oxide, expressed in mm. ; P,,, the initial pressure of azomethane; Po,the initial total pressure; P,, the minimum pressure attained during the experiment (when a minimum is actually observed); Pr, the final pressure; AP,, the pressure change a t minimum pressure (i. e., P, - Po). vaalad. is the total calculated volume of gas, corresponding to Pr as given in the table, reduced to 0” and 760 mm. and expressed in cc.; Vobsd. is the actually observed volume reduced to the same conditions; Vu is the volume under the same conditions of the dry undissolved product gases after bubbling through acid and alkaline solutions as already described. The significance of AP,/PNo is obvious. N2/NO represents the ratio of volume of dry nitrogen produced to volume of nitric oxide originally present, the latter being calculated to the conditions of (15) Burton, THISJ O U R N A L , 68, 1645 (193R>.

12

THOMAS W. DAVIS,FRANCIS P. JAHN AND MILTONBURTON

svMMARY OF

Experiment Temp., "C.

Ptio

P., P O

PUl Pi

-APm

cH4, %

% CaHc, % Nz, % C2&,

Total M o r Dh I/ealcd.

Vohd.

V"

Vol. 60

TABLEI ANALYTICAL RESULTS'

18" 20 10.1 39.7 49.8 41.7 42.7 8.1 0.7 11.8 0.9 86.9 100.3

16 20 9.9 100.5 110.4 102.0 103.9 8.4 0 1.9 4.7 93.8 100.4

17 20 10.2 102.2 112.4 104.5 112.3 7.9 3.7 28.4 0.5 67.5 loo. 1

lBd

14

20 10.0 102.9 112.9 106.8 107.0 6.1 4.8 8.3 0 86.7 99.8

20

13.3 74.8 88.1 77.3 77.6 10.8 0 0.8 1.3 97.8 99.9

D 53.2 52.7 9.2 80 64 58

D 129.3 131.3 9.5 85 72 63

D 140.0 141.7 29.1 77 152 64

D 133.2 133.8 10.4 61 83 81

95.9 96.7 10.0 81 59 58

M

IO* 78 13.4 97.7 llI.l 109.5 111.8 1.6 0 ?

0 97.3

... M

116.2 106.7 11.2 12 78

11 78 20.5 103.3 123.8

12 162 13.6 103.5 117.1

13 222 13.3 105.1 118.4

120.0

118.7

120.8

0 0.9 .6 98.3 99.8

0 2.9 0.5 96.5 99.9

D 124.8 117.5 10.0

99.5 92.3 8.8

...

...

*..

D

... ...

...

... 3.7 2.7 0 93.7 100.1

D 89.0 82.5 8.5

... ...

... -APm/Ptio X 10' 74 46 80 Nn/NO X 10' ... (N%/NO)mX 10' The percentages are on the basis of gas not soluble in hydrochloric acid or stannous chloride. All symbols are Ne by actual measurement = M; by difference = D. ' In expt. 18 the prodexplained under Procedure end Results. ucts were heated for thirty-five minutes a t 100". The pressure at 20' after this heating was 51.3 mm. corresponding In expt. 10 the pressure continued to increase Expt. 19 was conducted with unfiltered light. t o v&d., 64 cc. after the arc was extinguished; PIgiven is the last observed value before removal of the gas for analysis. Pfat the conclusion of illumination was actually 110.9 mm. corresponding t o Vmied., 115.3 cc. measurement of the volume of nitrogen. (Nz/NO)m is the ratio (calculated from the data) at the time when the pressure was a minimum. This was done simply by making the rough assumption that practically no alkane gas was produced M o r e minimum pressure was attained. From the volume of alkane gas given off in the run the volume of nitrogen produced after attainment of the minimum was cakulated on the basis of previously published data.@The general validity of the method is shown by the fairly good agreement of experiments 14, 16, and 1-7. The much higher value of (N2/NO)m in experiment 19,conducted with unfiltered light, is presumably real. Attention should be called t o the fact that, although there was no indication of the formation of any solid or liquid product at 78" and above, deposition of some product took place at 20". The amount of such deposition was almost imperceptible; however, when no filter was employed a fog appeared almost immediately during the photolysis. After the completion of this last experiment, the residue formed by the fog was examined. When the reaction vessel was opened t o the air the residue appeared t o diminish in amount; on examination under a hand lens it appeared to be made of many fine droplets of colorless liqnid. The flask was set over a dry-ice container in order to draw the liquid into a drop which might be examined. After six hours of this trkatment the residue had taken the form of a iiIm and now appeared brown, oily and viscous. The vapors were tested before this treatment and found not alkaline-the odor was that of azomethane without olfactory evidence of formaldehyde. After the six-hour period the odor was practically im-

...

perceptible. Some of the residue was swabbed up on a piece of filter paper; a qualitative nitrogen test run by Mr. W m e r H. Bromund was positive. It also should be mentioned that a white deposit appeared in the Tapler pump and traps in the early experiments and increased in amount as the investigation proceeded. When this deposit was exposed to the air no change was observed.

Discussion

The Nature of the RplTO Compounds.-The normal course of the photolysis of azomethane involves a pressure increase.6 Table I shows that when nitric oxide is present at 20' a pressure decrease occurs. The usual assumption11*12is that a reaction takes place between nitric oxide and any free radical, although there is as yet no definite information in the literature as to the precise nature of the compounds. The conclusions to be derived concerning the mechanism depend largely on what we may say concerning the vo€atility of these RNO compounds. There are at least three possible reactions by which nitric oxide may disappear from the reaction mixture. Considering what we know of the photolysis of azomethane, none of the following appear utterly improbable.

Jan., 1938

+ + +

NITRICOXIDEIN

CHC NO +CHaNO NO ---f CHsN-NNO CW=N CE4A NO --f CHANO

THE

13

(16)

The Effect of Wave Length.-Apparently,

(17) (18)

whatever the products are, they are stable (apart from possible polymerization) to light of 3500 A. This may be concluded particularly from experiment 17 which was continued past the point of minimum pressure until the pressure had returned to approximately its original value. Burton, Davis and Taylors have indicated that in the initial stages of the photolysis of azomethane, at the approximate temperature and pressure of experiment 17, the ratio of nitrogen to carbon atoms in the product is -1.25, when only the insoluble products are considered. When the amount of nitrogen formed after minimum pressure is calculated on this basis and the value of (N,/NO), is calculated from the result, it is found that the value so obtained agrees very well with similar ones for experiments 14 and 16. It should be noted that the agreement between V,..lcd. and Vobsd. in this experiment is good, indicating that no condensation of products took place during compression prior to analysis. Consequently, it appears that any RNO product formed during the process of pressure decrease is not decomposing or dissociating (or otherwise interfering with the normal course of the photolysis) during the period of pressure increase, when presumably all the nitric oxide has been used up, and that the RNO compounds are stable at 3500 A. This is not surprising in view of the results of Thompson and Linnett.'lb It is more difficult to explain the results with unfiltered light. Burton, Davis and Taylor6have already shown that the use of unfiltered light is practically without effect on the nitrogen yield in the photolysis of pure azomethane. In experiment 19 the nitrogen yield was much higher, although -Urnwas smaller, than in the experiments with filtered light a t 20'; evidently, the result 'is traceable to the effect of nitric oxide. The same phenomenon is probably responsible for both observations; e. g., the formation of an RNO product causes pressure decrease, the decomposition of such a product (or dissociation of its polymer) causes a pressure increase. Light of wave length between 3500 and 2900 A. is apparently sufllcient to effect such a decomposition and thus to cause the pressure decrease to be less than would be the case with filtered light. This is not contradictory to the results of Thompson and Linnett;'lb it merely indicates that it is probably a different RNO compound with which we are

It should be noted that there is no attempt to elucidate the nature of these products. CHsNO is isomeric with formaldoxime, b. p. 84', and with formamide, b. p. 210.7'. Whether an RNO product formed actually isomerizes to one of these compounds cannot be stated. However, the known properties of nitroso compounds1ea would indicate a ready isomerization of CHaNO to CH2=NOH although the reverse reaction would not be expected to occur. There is also the possibility of formation of a polymer; formaldoxime forms a termolecular polymer with great ease.lsb A satisfactory interpretation of the work of Thompson and Linnett,llb who studied the photolysis of a mixture of mercury dimethyl and nitric oxide, requires that CHdNObe formed and that it or its isomer subsequently precipitated. Furthermore, the CHsNO or its isomer must have been stable to light of -2537 A. As to the properties of the other hypothetical products shown in reactions 17 and 18, we are without definite information. The Stability of the Products.-It has been mentioned that when the reaction mixture was permitted to stand after the completion of experiment 10 at 78O, the pressure slowly increased, indicating the decomposition of one of the products of the reaction. No such effect was noted at 20', indicating the stability of the products at that temperature. When, as in experiment 18, the products were heated to > 100' for thirty-five minutes and allowed to cool to the original temperature, a distinct pressure increase was noted. In fact, the pressure attained was higher than the initial pressure. However, the volume of gas as measured over mercury corresponded not to that pressure but to the pressure immediately after the completion of the photolysis and before the heating operation. Obviously, most of the products formed by decomposition were condensed out in the Topler pump.'' The depolymerization of condensed RNO compounds would account for most of this result. The RNO molecule itself apparently decomposes but slightly, if a t all; the total volume of alkane gas in this experiment was 1.2 cc. The signScance of the values for Nr/NO and (Ns/NO), will be discussed later. (16) Sidgwick, Taylor and Baker, "The Organic Chemistry of Nitrogen Compounds," Oxford University Press, Oxford, 1937, (a) pp. 168,204, (b) p. 172, (c) p. 167. (17) I t should be noted that all the analyses were conducted at

about 30'.

PHOTOLYSIS OF AZOMETHANE

1.1

l r - t o ~ aW. s DAVIS,FRANCIS P. JAHN

dealing in this case. Inasmuch as nitrogen appears to be produced in this photodecomposition it seems likely that the compound is either -AP,, in CHIN=NNO or (CH&NNC&NO. this experiment was 6.1 mm.; in other similar experiments a t 20’ the ax-erage value of -AP, was 8.1 mm. If we assume that the difference of 2 mm. is caused by the formation of nitrogen as a result of RNO decomposition, we could calculate an “ideal” value of (N2/NO),, not including the amount of nitrogen so formed, of 0.60. This is slightly below the value of 0.83 for experiments lli and 17 and indicates that part of the 2 mm. difference is probably attributable to the formation of some alkane gas, but a smaller amount than is usually associated with the formation of nitrogen.ls If, instead of calculating on the basis of the -A P,, variation, the figure fur (Nz/NO), had been computed on the basis of the usual nitrogen-carbon ratio,6as was done in experiment 17, then the value would have been -0.69, which is much higher than the usual average. ‘There consequently appears to be little doubt that the excess nitrogen originates in an RNO compound and not in azomethane itself. Judging from the work of Thompson and Linnett, this cannot be CHdNO. It will be shown below, in the discussion of the path of reaction, that the formation of CHBNNNO is improbable. It may therefore be inferred that CH3AN0 is the R S O compound involved. The Effect of Concentration.-In experiments 18, 16, 17 and 14 a t 20’ the ratios of the initial concentrations of nitric oxide and azomethane were varied. The greater the Pxo/Paz0ratio, the greater would be the chance of reaction 16 occurring before reaction 6. I t is evident from Table I that the values of (Nz/SO), decrease I“ slightly with increase of PxuiPnLo. Inasmuch as the nitric oxide : azomethane ratio decreases during the photolysis a great variation in (Na/ NO),, could not be expected. The significance of the variation is discussed later. The Effect of Temperature.--It has been shown above that increase of temperature is (18) It can be shown by a simple calculation that in order for the “ideal” value for ( N Z / N O )to ~ be 0.83,the ratio of nitrogen and carbon atoms in the gas produced in the decomposition of the Rh’0 compound would have to be -5.22. (19) In experiment 18 ( N z , ” ~ ) , ~was calculated on the assumption that all the nitrogen originated during the photolysis; the amount produced during the period of pressure increase was then calculated on the basis of the data of Burton, Davis and Taylor’ and the value for (NdNO)m determined to be 0.58. Had it been calculated on assumptions similar to those involved in the similar calculation of experimnt 19 (see above), it would have been found to he -0.51. Evidently t h e value of 0 58 is a maximum estimate.

AND

MILTON BURTOK

Vol. (i0

probably effective in dissociating an RNO polymer but that it seems to have but a slight further effect in decomposing the RNO molecule itself. It was noted in experiment 18 that the observed increase in pressure at 20°, after heating the products of the photolysis, was unrelated to the volume as measured over mercury. It was assumed consequently that the RNO, dissociated from the polymer, reprecipitated a t atmospheric pressures. Obviously the decrease of volume noted in this way should not exceed the initial volume of nitric oxide, TINo, corrected to 0 ” and 760 mm. In Table I1 the differences in the calculated and observed volumes are compared with VND for five experiments. It is clear that in none of the experiments did Vcalcd. - Vobsd. approach the maximum allowable according to this hypothesis. The nearest approach was in the case of experiment 18 where the products were heated after the completion of the photolysis. TABLEI1 CORRELATION OF VOLUMEDATA Expt. 18‘ 10 11 12 13 78 78 Temp.,’C. ... 163 322 7.3 7.2 t’oelod.-vobd. 11 3 9 5 6.5 VN0 12.6 14.0 21 3 11 4 9 8 a The value of Vanlcd. used in expt. 18 is that corresponding to a pressure of 61.3 mm. See Table I, note G.

It may be noted from Table I that increase of temperature is without considerable effect on the nature of the products. The comparison is best made among experiments 14, 11, 12 and 13, covering the range 20-222’. In experiment 14 the minimum point was just passed a t the completion of the run ; i. e., the nitric oxide was just consumed. In the other experiments the time of the run was presumably insufficient to consume all the nitric oxide. It may be seen that a t the lower temperatures practically no alkane is produced as long as any nitric oxide may be presumed to be present. The amount of alkane produced, although small, increased steadily a t 168 and 222’. Referring to Fig. 1, it may be seen that a t the lower temperature (20’) there is a rapid decrease of pressure in the initial stages. At higher temperatures (?So) this rate decreases. At 163’ the initial pressure increase may be due to a slight Draper effect; however, there is a small but definite pressure increase throughout the run. At 222’ the rate of pressure increase is somewhat greater. The Path of the Reaction. Path 1.-The failure to observe an appreciable amount of alkane

Jan., 1935

NITRICOXIDEIN

THE

PHOTOLYSIS OF AZOMETHANE

15

b.-Assume CH3ANO to be volatile. If the gas in the early stages of the photolysis definitely excludes reaction 1 as a possible primary process. path be by way of reactions 2 and 6 followed Path 2.-Assume CHaNO completely involatile alternatively by 18 or the competing reaction CH3A 4-NO +Nz (CH&NO (18a) or polymerized to an involatile compound, as seems more likely. If the path of the photolysis the pressure changes and the N2,JN0 ratio (where and subsequent steps be exclusively by reactions N2,trepresents the volume of nitrogen produced 2 and 16, a decrease of pressure would be observed. in the time required to consume all the nitric -AP, would exactly equal PNo. If the photoly- oxide) would be related to the temperature. If sis be continued for exactly the time required to it be assumed that the energy of activation of consume all the nitric oxide, then N2/N0 should reaction 18a is higher than that of 18, low tembe 0.5. Neither the effect of temperature on peratures favor 18 and high temperatures favor pressure change nor the magnitude of the N2/NO 18a. At sufficiently low temperatures, the ratios is consistent with the hypothesis that this N2,JN0 ratio would be 0.5 and the maximum pressure decrease would equal PNo. As the path constitutes the exclusive mechanism. Path 3. a.-If the path were by reactions 3, temperature is raised, reaction 18a would be16 and 17 there would be no nitrogen produced. come more rapid; the yield of nitrogen would This is contrary to the observations. The possi- increase so that N2,JN0 becomes greater than bility that the paths 2 and 3 occur simultaneously 0.5. Simultaneously, the ratio - AP,/PN, is immediately ruled out by the fact that the would decrease to a value less than 1.0. At a N2/NO ratios observed greatly exceed any value sufficiently high temperature all the CHsA may expected on the basis of this assumption. react according to reaction 18a. However, at b.-The assumption that the CHINNNO such a temperature reaction 20 would also be formed in reaction 17 subsequently decomposes, expected.

+

CHINNNO -+ CHINO

+ Nf

(17a)

yielding nitrogen, is by itself inadequate to explain the results obtained. The assumption of immediate decomposition (i. e., merely momentary existence) of CHsNNNO is formally equivalent to path 2. Thus, if path 2 (as will appear later) is an adequate representation of part of the reaction, path 3b may likewise be. Since nothing is known of the properties of the hypothetical CHsNNNO, there can be no definite conclusion on this point. 6.-There is the possibility that CHaNNNO is not formed at all but that an alternative reaction occurs

(CHs)aNO +HCHO f (CH&NH

(20)

Thus, at high temperatures N2,t/N0 may approach 1.5 and the pressure would actually increase an amount equal to PNo in the same time. At intermediate temperatures (assuming equal probabilities of reactions 18 and 18a), N2,JN0 would be 1.0 and there would be no initial pressure change. At 163 and 222' the pressure increases while Nz/NO is still less than 1.0. This path cannot, therefore, be the exclusive mechanism. c.-Assume C H d N O to be involatile. As before, the N2,t/N0 ratio would be 0.5 a t very AP,/PNo would then be low temperatures. -2.0. However, at high temperatures the beCHaNN NO +CHaNO Nz havior would be similar to that of path 4b. At (17b) This path also is formally equivalent to path 2. an intermediate temperature corresponding to It is quite evident that it cannot be said that N2,t/N0 = 1.0 there would be a pressure decrease reaction 3 does not occur. It can only be stated unless the CHANO volatilized, which case has that, if it does, the subsequent reactions are been considered above as path 4b. Thus 4c such as to conceal that fact. cannot be the exclusive mechanism. Path 4. a.-Assume CHANO to be volatile. Combination: Paths 2, 4b and 4c.-The asThe path by way of reactions 2, 6, 18 and sumption here is that azomethane decomposes CHaANO +Nz (CHa)aNO exclusively by reaction 2 and that the methyl (19) with the rate of the last reaction dependent on radicals may then disappear via reactions 6 temperature, is inadmissible. Experiment 18, and 16 forming CH3A and CH3NO. The former when compared with the other experiments a t would react via 18 or 18a to give CH&NO or 20°, clearly shows that if CHANO is formed it Nz and (CH3)3N0. CH3N0 probably polycertainly does not undergo reaction 19 a t 100' to merizes during the photolysis as already menany measurable extent. tioned; a t the higher temperatures the polymer

+

+

+

IC,

w.

THOMAS DAVIS,FRANCIS P.JAHN AND MILTONBURTON

probably dissociates, possibly forming formaldoxime which would be volatile and might itself dissociate to some

4HzO + eo

YHCX CHsNO

(21)

As for CHaANO it may or may not be volatile at room temperature. Evidently, there are not enough data with which to decide the precise course of the reaction steps, particularly since the fate of the RNO products can only be roughly surmised. Combination Including Path ,?.-The substitution of path 3a for path 2 in the above combination would result in a nitrogen ddciency unless we assume a t the same time a high probability of reaction 18a in comparison With reaction 18. However, this seems unlikely if reaction 17 is faster than 17b, as it would have to be were path 3a to occur. It thus appears that, if any of the reaction is by path 3a, it is but a very small fraction, On the other hand, paths 3b or 3c may be substituted for path 2 in the above combination without adding any new hypotheses or conclusions ; no dficulties are introduced and none are explained. The assumption of such paths may be correct but it is not necessary. It seems that the most reasonable assumptioris as to the course of the reaction are: 1. Reaction 1 occurs to a very limited extent if at d. 2. Reaction 2 does occur. If reaction 3 takes place a t all it is either followed so rapidly and so inevitably by reaction 22 CHaNN

+ Nz

CHB

(22)

Vol. 60

and Taylor6 at 20’ indicate that one out of every five methyl radicals formed in the primary step of the photolysis of azomethane disappears permanently through the step 6. The other four radicals ultimately appear as ethane molecules. The ethane cannot be formed by step 15 for, although no ethane is formed in the presence of nitric oxide, reaction 6 appears to take place nevertheless and is evidently a much faster reaction than 15. The production of ethane in such large quantities in the photolysis of pure azomethane must therefore be attributed to some subsequent reaction of the radical CHSA or a compound formed from that radical. Such an assumption might also account for the production of methane and propane already reported.”’ Effect of the NO Product on the Course of the Photolysis.-In considering the effect of wave length it has been shown that the product formed with nitric oxide does not interfere with the normal course of the photolysis. Consequently, the coiiclusions at which we arrive, based on the assumption that nitric oxide (and not any RNO compound) alone reacts with the products of the photolysis of azomethane, are not unreasonable. Conclusion.-Work of previous investigators founded on the assumption that azomethane decomposes almost exclusively into free methyl radicals and molecular nitrogen appears to be justified by OUT photochemical results. There is a possibility, however, that their calculations may be somewhat distorted because of the fact that methyl radicals appear to combine rather readily with azomethane molecules. Complex molecules so formed appear to be the source of ethane and other alkane gas noted in the photolysis of azomethane. It seems likely that practically no ethane is produced by the direct combination of methyl radicals formed during the photolysis of azomethane.

or so well masked by subsequent reactions that it cannot be distinguished from reaction 2 . 3. Methyl radicals disappear either by reaction 16 or by reaction 6 followed by 18 or 18a. The relative amounts of the reactions 16 and 6 and the ultimate fate of the products cannot be determined with the data at hand. summary It should be noted that the variation of (Nz,’ 1. Analyses are reported of the products formed S O ) , with concentration is in agreement with this type of path. Evidently, the greater the by the photolysis of azomethane in the presence of relative amount of nitric oxide, the greater the nitric oxide. 2 . When nitric oxide is present during the phoprobability of reaction 16 in comparison with reactions 6 and 18 or 18a. This would decrease tolysis of azomethane, practically no alkane gas is the amount of nitrogen produced by reaction lea, produced. 3. The photolysis of azomethane seems to proin line with the results reported. The Source of Ethane in the Photolysis of ceed almost exclusively by a path involving the Ammethane.-The experiments of Burton, Davis formation of free methyl radicals and nitrogen (20) Dunstan and &w.i, J Chein Soc., 73, 353 (1898) molecules in the primary act.

FREEZING POINTS OF SOMEPUREHYDROCARBONS

Jan., 1938

4. The methyl radicals subsequently combine with azomethane molecules yielding more complicated radicals and molecules. 5. The ethane and other alkane gas produced during the photolysis of azomethane results from a decomposition reaction involving the more complicated radicals and molecules. Practically none of the ethane formed results from a combination of the methyl radicals; methyl radicals do not combine readily a t ordinary temperatures to form ethane.

[CONTRIBUTION OF

THE

17

6. Some compound formed by the interaction of nitric oxide and the products of the photolysis of azomethane is thermally unstable, yielding mainly substances which are soluble in acid solution. 7. Some compound so formed is more stable to light than is azomethane. When it is decomposed by light, nitrogen and a small amount of alkane gas are among the products. UNIVERSITY HEIGIFTS NEW YORX, N. Y . BERKELEY,CALIF.

RESEARCH LABORATORY OF

THE

RECEIVED OCTOBBR 11, 1937

BATAAFSCHE PETROLEUM MAATSCEtAPPIJ]

Freezing Points of a Number of Pure Hydrocarbons of the Gasoline Boiling Range and of Some of their Binary Mixtures BY J. SMITTENBERG, H. HOOGAND R. A. HENKES The freezing point or melting point is one of the most useful physical properties in the identification of chemical substances and in the veriiication of their degree of purity. For this reason we started an investigation of the freezing points of a number of hydrocarbons of the gasoline boiling range, which were available in our laboratory. Since it was desirable to have some idea of the effect of impurities upon the behavior of the hydrocarbons when freezing, a number of binary hydrocarbon mixtures were also tentatively examined. 1. Preparation and Purification of the Hydrocarbons.Nearly all the hydrocarbons were prepared synthetically, some in our laboratory and some in the laboratory for organic chemistry of the University of Amsterdam under the guidance of Professor Dr. J. P. Wibaut. As a rule the hydrocarbons were purified by sharp rectification with a fractionating column of about 22 theoretical plates. Each preparation was separated into at least ten fractions and only those fractions whose boiling points differed by less than 0.2’ and whose refractive indices differed by less than 0.0002 were accepted as “pure.” The preparation and examination of these hydrocarbons will be described shortly in extenso in a separate article. The provisional results of the determinations of the boiling point, the refractive index and the freezing point and/or melting point are given in Table I. A number of other physical properties are at present being determined. 2. Apparatus and Methods for the Determination of Freezing and Melting Points.-A platinum resistance thermometer, consisting of about 85 cm. of platinum wire 0.033 mm. in diameter and made to De Leeuw’s pattern,’ was used for most of the determinations. The total (1) De Leeuw, 2. physik. Ckcm., 77, 304 (1911).

length of this thermometer, S, was 20 cm., its diameter 0.8 an.,and the volume of the actual measuring part, 1.5 cc. I n some instances a larger thermometer, L, was used, consisting of 100 cm.of platinum wire 0.05 mm. in diameter. The total length of this thermometer was 50 cm., its diameter 1.2 cm.and the volume of the measuring part 5 cc. A Wheatstone bridge was used to measure the resistance; the values of the deflections of the galvanometer were calibrated with the aid of standard resistances, which could be placed in the bridge instead of the thermometer. Four or five fixed points were used for the temperature standardization of the resistance thermometers, as follows Fixed point

no.

Substanre

Accepted temp., ‘ C .

1 2 3 4 5

Ice Mercury #-Heptane n-Pentane Oxygen

M.P. 0.00 F. p. - 38.87 F. p. - 90.8 F. p. -129.73 B. p. -183.0

Resistance of thermometer in ohms S L

99.06 84.85 65.65 50.82 30.7

45.38 38.45 29.08 21.87

The n-heptane was a commercial product, obtained from the California Chem. Co., the freezing point of which is guaranteed by the National Bureau of Standards t o be -90.8’. The n-pentane was a synthetic product which, according to its physical constants and cooling curve, was very pure. The value given by Mair’ was accepted for its freezing point. The following procedure was adopted for rapid conversion of intermediate resistances to the corresponding temperatures. For each thermometer the constants a and b of the formula t=aR-b (1) were calculated from the fixed points 2 and 3. In this formula t represents the fixed temperature in degrees centi(2) B. J. Mair, Bur. Standards J . Rcscarck, 9, 457 (1932).